Coalition formation via negotiation in multiagent systems with voluntary attacks

Argumentation networks are put forward by Dung considering only one kind of attack among arguments. In this paper, we propose to extend Dung's argumentation framework with voluntary attacks in the context of multiagent systems, characterized by the possibility of the attacker to decide whether to attack or not. Enabling voluntary attacks impacts on the acceptability of the arguments in the framework, and therefore it becomes subject of debate between the agents. Agents can negotiate about which subset of voluntary attacks can be raised, and they form coalitions after the negotiation process.